<p>A recently discovered instance of child sexual abuse material within the LAION-5B dataset has sparked controversy and concerns over responsible AI image training. LAION, a project aimed at developing an open-source dataset for image-to-text diffusion models, has temporarily taken down the datasets in response. This is not the first time LAION-5B has faced scrutiny, as it previously received criticism from renowned AI pioneer Andrew Ng. The dataset, which was found to contain at least 1,679 illegal images, was compiled by scraping social media posts and adult websites. In light of this discovery, LAION and Stability AI have taken immediate measures to remove the offending content and fine-tune their models for enhanced safety. The incident has also prompted US attorneys general to call for an investigation into the impact of AI on child exploitation. The findings highlight the crucial need for strict regulations, thorough content screening, and continuous oversight in the development of AI training datasets to prevent the dissemination and potential victimization associated with illegal and harmful imagery.</p>


Collections

An open-source AI dataset, LAION-5B, has been found to contain child sexual abuse material, leading to controversy and concerns over responsible AI training. The dataset has been temporarily taken down and measures are being taken to remove the offending content and enhance safety.

Controversial Open Source AI Dataset LAION-5B Found to Include Child Sexual Abuse Material

<p>Just another example of how webscraping is unethical and problematic</p>
